Leveraging long short term memory in air pollution prediction in Nairobi

Abstract

Air pollution poses a major environmental health risk, leading to approximately 6.7 million premature deaths annually. Key pollutants like PM2.5, carbon monoxide (CO), sulfur dioxide (SO₂), and ozone (O₃) significantly affect air quality. This study utilizes a Long Short-Term Memory (LSTM) deep neural network algorithm to predict air pollution levels, focusing on PM2.5 concentrations in Nairobi. Sensor data from GeoHealth Hub was split into training, validation, and testing datasets. The LSTM model, optimized with the Adam algorithm and evaluated using Root Mean Squared Error (RMSE), demonstrated superior accuracy over baseline models, offering valuable insights for future air quality management and mitigation efforts.